Comparison of Direct Syringe Filtration and Membrane Filtration for the Selective Isolation of Campylobacter jejuni from Ready-to-Eat Sprouts

Introduction

C ampylobacter spp. are microaerophilic bacteria that are known to cause food poisoning worldwide (Park et al., 1992). Among all Campylobacter species, Campylobacter jejuni causes 90% of foodborne campylobacteriosis (Brandl et al., 2004). Although poultry products and contaminated water are the primary sources of Campylobacter, fresh vegetables have become a significant source of campylobacteriosis (Kärenlampi and Hänninen, 1986; Evans et al., 2003; Brandl et al., 2004; Chai et al., 2007, 2008). Even though enrichment step in culture method enables the selective proliferation of target bacteria that pose public health risk, it allows the overgrowth of resistant bacteria making isolation even more difficult (Hyeon et al., 2009; Chon et al., 2011). The potential risk of C. jejuni by low reliability of the detection method could be connecting with the defect of the culture method. The isolation of C. jejuni by culture method from fresh produce is much more difficult than other types of foods because of its hostile environments and indigenous flora (Brandl et al., 2004; Chon et al., 2011). In particular, the level of indigenous flora is extremely high and diverse in ready-to-eat (RTE) sprouts compared with other types of RTE vegetables such as vegetable salad and fresh cut (Hyeon et al., 2009; Lee et al., 2010; Chon et al., 2011, 2012). The optimal conditions for germination of sprouts are ideal for bacterial proliferation and the high level of diverse indigenous flora can prevent the isolation of pathogenic bacteria (Loui et al., 2008; Chon et al., 2012; Ding et al., 2013). Consumers could be susceptible to food poisoning if the sprouts are contaminated with foodborne pathogens (Sikin et al., 2013). Thus, a more sensitive and selective culture method is necessary to detect C. jejuni in RTE sprouts.

To overcome the disadvantage of culture method, the direct application of membrane filters to the surface of the plating media has been previously used to exclude the unwanted microflora for the selective isolation of Campylobacter (Steele and McDermott, 1984; Kulkarni et al., 2002; Speegle et al., 2009; Jokinen et al., 2012). In our previous study, we reported that sample enrichment with a 0.65-μm pore size syringe filter might be a simple and useful option for increasing the recovery and selectivity rates (Chon et al., 2017). Few studies, however, have compared the use of syringe or membrane filtrations to culture method without filtration for the isolation of Campylobacter in RTE sprouts. In this study, we compared a culture method for the isolation of C. jejuni from sprout samples with a method that incorporated syringe or membrane filtrations.

Materials and Methods

Detection of C. jejuni in RTE sprout samples by culture method and two filtration methods

All experimental procedures for the detection of C. jejuni from RTE sprouts were conducted as described in the official isolation protocols of the Korea Food and Drug Administration (2014) with minor modification.

Each 25 g sample was put in a sterilized stomacher bag with 100 mL of Bolton broth (Oxoid, Hampshire, United Kingdom). After stomaching for 30 s, homogenated samples were pre-enriched in Bolton broth at 37°C for 4 h, and further enriched at 42°C for 44 h. All Bolton broth enrichments were processed following three separate pathways, culture method (unfiltered), membrane filtration, and syringe filtration method, as described in Figure 1. In the culture method without filtration, the unfiltered enrichment culture was inoculated onto a modified charcoal-cefoperazone-deoxycholate agar (mCCDA; Oxoid) plate by streaking with 10 μL sized disposable loop. For membrane filtration, five 20-μL drops (100 μL) of culture were inoculated on five area of a 0.65-μm pore size membrane filter (nitrocellulose membrane; Millipore, Schwalbach, Germany), as previously described (Speegle et al., 2009). The membrane filter was left in contact with the surface of a plate for 20 min. For syringe filtration, about 5 mL of each dilution was transferred into the syringe connected with a 0.65-μm pore size syringe filter (Minisart^®; Sartorius, Goettingen, Germany). The syringe plunger was depressed immediately, after which each filtrate was collected in conical tubes, and streaked with 10 μL sized disposable loop on a mCCDA plate. All plates were incubated at 42°C for 48 h microaerobically. Suspected colonies were removed and subcultured, and identification of C. jejuni was confirmed by colony polymerase chain reaction, as previously described (Denis et al., 1999).

OPEN IN VIEWER

FIG. 1.

A proposed protocol for Campylobacter jejuni isolation from ready-to-eat sprouts. To generate partial positives and partial negatives for statistical comparison between 20 subsamples, C. jejuni NCTC 11168 was artificially inoculated into 500 g bulk samples. Comparison of competing flora on agar plates inoculated with unfiltered method, syringe-filtered, or membrane-filtered enrichments. mCCD, modified charcoal-cefoperazone-deoxycholate.

Results and Discussion

The performances of the culture methods with or without filtration for the isolation of C. jejuni from four types of RTE sprout samples are compared in Table 1. Both filtration methods resulted in a significantly higher (p < 0.05) isolation rate of Campylobacter positives (58–61%) than the culture method (10%) without filtration. Although no statistical difference was observed (p > 0.05), membrane filtration (61%) yielded the highest positives compared with syringe filtration (58%; Table 1).

Comparison of competing flora on agar plates inoculated with unfiltered, syringe-filtered, or membrane-filtered enrichment is presented in Table 2. In unfiltered samples, most agar plates (98%) were contaminated with competing flora. In most unfiltered samples, high competing flora covered most of the plate, making differentiation and picking of suspicious colonies difficult. However, much less plates were contaminated with competing flora in both filtration methods. With the syringe filtration method, only 5% of plates were contaminated, whereas no competing flora was observed with the membrane filtration method (0%, Table 2). The number of competing flora on plates from syringe filtration was low and did not mask the suspicious colonies. Some competing flora on the mCCDA plates from syringe filtration were biochemically confirmed by using a VITEK 2 assay. Most strains were identified as Sphingomonas paucimobilis, followed by Aeromonas salmonicida (data not shown).

We previously reported that a combination of enrichment and syringe filtration results in a significantly higher isolation rate and minimal contamination of Campylobacter from chicken carcasses (Chon et al., 2017). However, few studies have evaluated the efficacy of filtration methods in fresh produce samples. We found that membrane filtration method yields the highest isolation rate of C. jejuni among tested methods (Table 1), even though no statistical difference was found with syringe filtration method (p > 0.05). The selectivity was also better in the membrane filtration method compared with other tested methods. No contaminated plate with competing flora was observed in the membrane filtration method, whereas 5% of plates were contaminated with competing flora when syringe filtration was used (Table 2). As described earlier, membrane filtration has been commonly used to exclude unwanted microflora during the selective isolation of Campylobacter. However, membrane filtration is associated with several technical difficulties, such as a lengthy filtration time. Campylobacter cells inoculated on a membrane filter covering the surface of plating media are allowed to pass through the membrane for 15–20 min (Steele and McDermott, 1984; Speegle et al., 2009). Furthermore, contamination of the plate can arise as a result of filter mishandling, which allows unfiltered liquid to come into contact with the agar surface (Speegle et al., 2009; Chon et al., 2017). However, syringe filtration saves time as the culture broth being immediately filtered in a few seconds right after the enrichment step. The possibility of contamination by mishandling is also much lower in the syringe filtration method as it does not involve direct handling of the filter paper.

Although cell motility and gravitation have been regarded as an essential driving force for the passage of cells through membrane filters (Steele and McDermott, 1984; Wilson and Aitchison, 2007; Speegle et al., 2009), positive pushing pressure appears to play a critical role in syringe filtration (Chon et al., 2017). The charge of the cell membrane is also a determining factor of target bacteria filtration. Official Campylobacter isolation protocol from U.S. Food and Drug Administration is currently adopting the use of 45 μm Zetapor filters for water sample. The filters have a positive charge, making the negative-charged Gram-negative bacteria are more effectively trapped in the filter (Hunt et al., 2001).

Our previous study was the first to report that syringe filtration method could be employed for Campylobacter isolation (Chon et al., 2017). In this study, we found that syringe filtration is still a reliable method for the isolation of C. jejuni in RTE sprouts. Enrichment incorporating syringe filtration may be a simple and useful option to increase recovery and selectivity rates and to decrease contamination and procedure time during C. jejuni isolation from fresh produce. However, the information generated by this study is limited because inoculation level between the syringe (10 μL, streaking) and membrane filtrations (100 μL) according to the methods described in previous studies we referred to (Speegle et al., 2009; Chon et al., 2017). Even though this difference may not have significant effect on the isolation rate for qualitative detection as presented in this study, it should be considered as an important factor for enumeration of C. jejuni in food samples.

The direct application of membrane filters to the surface of the plating media has been previously used to exclude the unwanted microflora for the selective isolation of Campylobacter. Campylobacters may be obtained from the top of membrane, or from the filtrates. For example, Fluoropore and Durapore filters with 0.2–0.5 μm pore size effectively retained campylobacter cells from water samples, whereas cellulose ester filters with 0.22–0.45 μm pore size did not (Oyofo and Rollins, 1993) unless the elution method was modified (Tissier et al., 2012). Even though many previous studies compared and evaluated the efficacy of filtration in clinical and water sample (Steele and McDermott, 1984; Bolton et al., 1988; Wilson and Aitchison, 2007; Jokinen et al., 2012), few studies have been conducted in food sample. Application of filtration in food samples is a bit different from other samples. The efficacy of the filter highly depends on the food matrix rather than size of bacteria, because food particles more easily clog the pores than bacteria. Thus, the application of filtration should be evaluated in foods on different experimental conditions. In the future study, the sensitivity and selectivity of various membrane filters with different membrane types and pore size should be evaluated in poultry meat and vegetable samples.